February 16, 20251 yr My server only has a runtime of 10 days or so, and prior to that it was having quite a few problems freezing up which required reboots when I was doing massive backups over LAN to another Unraid server. I see a few things in my syslog that are repeated quite regularly that might be the cause of it, but I don't know what they mean, if they're the issue, nor how to fix them. They're repeated messages about various ethernet ports being disabled, entering a blocking state, entering a forwarding state, etc. There were also some strings of errors when one of my tailscale mounted remotes in unassigned devices was down, but that is resolved now. I've attached my diagnostics and would greatly appreciate some insight/assistance getting to the bottom of this. unraidplex-diagnostics-20250216-0743.zip
February 17, 20251 yr Community Expert Main issue spamming the log are port multiplier issues with the add-on controller: Feb 15 06:43:08 UnRaidPlex kernel: ahci 0000:02:00.0: FBS is disabled Feb 15 06:43:08 UnRaidPlex kernel: ahci 0000:02:00.0: FBS is enabled Feb 15 06:43:08 UnRaidPlex kernel: ata1.00: SATA link up 6.0 Gbps (SStatus 133 SControl 330) Feb 15 06:43:08 UnRaidPlex kernel: ahci 0000:02:00.0: FBS is disabled Feb 15 06:43:08 UnRaidPlex kernel: ahci 0000:02:00.0: FBS is enabled Feb 15 06:43:08 UnRaidPlex kernel: ata1.01: SATA link up 6.0 Gbps (SStatus 133 SControl 330) Feb 15 06:43:08 UnRaidPlex kernel: ata1.00: configured for UDMA/133 Feb 15 06:43:08 UnRaidPlex kernel: ata1.01: configured for UDMA/133 Feb 15 10:23:57 UnRaidPlex kernel: ata1.00: failed to read SCR 1 (Emask=0x40) Feb 15 10:23:57 UnRaidPlex kernel: ata1.01: failed to read SCR 1 (Emask=0x40) Feb 15 10:23:58 UnRaidPlex kernel: ata1.15: SATA link up 6.0 Gbps (SStatus 133 SControl 300) Controllers with port SATA multiplier are not recommended, see here:
February 19, 20251 yr Author Hmm...the adapter I have is using a JMB585 chip which is one of the recommended chips listed on the page you referenced. I don't recall this being an issue in the past, so do you think the adapter is simply going bad? I have a spare NVME M.2 SATA adapter, so should I try swapping it out? This is the exact model I have: https://a.co/d/gVCZ8er Or should I just ditch the NVME M.2 SATA adapter and go with an HBA like this? https://a.co/d/aZH0HVN My concern with switching over to an HBA is that I read somewhere that disk location is part of dual parity calculation. Is it merely the disk layout in the Unraid GUI that needs to remain the same, or is it the physical location on the mobo that needs to remain the same? If all I need to do is take a screenshot of my disk locations then make sure they're still in the same slots in the GUI then I'm cool with that. --Please forgive my ignorance here. I currently have a GPU occupying the sole PCIe slot on my ASRock z690m-itx/ax mobo that I passthru to a Windows 11 VM, but I'm willing to ditch that if it will mean my data is safer and drives are more stable. I wonder if this is why I was experiencing system lockups a couple of weeks ago when I was attempting to backup about 25.5TB of data to a secondary Unraid server.
February 19, 20251 yr Community Expert Solution 7 hours ago, Sup3rTr00p said: the adapter I have is using a JMB585 chip which is one of the recommended chips listed on the page you referenced. Yes, but it's using a SATA port multiplier, do you have it connected to some sort of enclosure?
February 19, 20251 yr Author Ah HA!!! One of the SATA ports is populated with 2 x SATA m.2 drives inside a QNAP Dual M.2 SATA SSD to 2.5" SATA RAID Adapter Converter set to Individual Mode: (https://www.qnap.com/en/product/qda-a2mar). Looking over the documentation, it appears Individual Mode acts as a port multiplier, so that must be the issue! I am using that as a redundant cache drive but maybe I need to rethink how I'm doing this. I wonder if setting the enclosure to RAID 1 and having the redundancy at the hardware level rather than doing it in software would be better. I went ahead and ordered an HBA just in case, and if that would be the better route to go than the JMB535 M.2 adapter, that would free up several more SATA ports that I can use for separate 2.5" SATA drives which I can use for redundant cache pools. Thanks for helping me sort this out!
February 21, 20251 yr Author Quick follow up: I pulled the JMB535 m.2 SATA x5 adapter, removed my Nvidia Quadro K620 gpu, added an HBA, and added another NVME drive. I also changed the QNAP m.2 SATA enclosure to RAID1 mode and plugged it into one of the motherboard SATA ports. Upon firing it up, I simply had to format the QNAP SATA SSD enclosure, add it to a new cache pool, and add the new NVME drive to my existing cache pool. Now I have 2 protected SSD cache pools: the NVME one for /domains, /appdata, /system, and /nextcloud; and the SATA one for /data, /isos, and /syslogs. Everything appears to be running well and there's no more log flooding with port multiplier errors! The only thing I "lost" in the scenario is the GPU that I was passing through to my Windows 11 VM, but I rarely used that anyways. I did "gain" 3 SATA ports on the motherboard, so I may toss in some additional SATA SSD's for VM's in the near future. Thank you once again for all the help JorgeB!!!
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.